Planar Rogowski coils on PCB

Murata makes some cute little dc/dc converters where the windings are all traces and vias. There is a ferrite toroid buried *inside* the PC board.

A similar idea could make a current transformer.

Yes, there are modules like these, but it is still exotica. The Rogowski coil does not use a core (that's the whole idea), so can be implemented on standard PCB substrates. If the vendor can make a 4-layer PCB, they can make this as well, without asking any additional questions.

I am tempted to prototype it just out of curiosity.

Best regards, Piotr

On a sunny day (Wed, 2 Sep 2020 08:34:18 +0200) it happened Piotr Wyderski wrote in :

It is nice, first I looked at the PCB from the top and tought 'this cannot work', then I realized from the article it was a 3 D coil wound around the field lines.. Very nice!

Or buy a 3-cent current shunt!

For a hundred of amps? Interesting. Best regards, Piotr

Here is one more, likely by the guy from the movie:

Best regards, Piotr

Hi Piotr, please have a look here:

some experiments with promising results were performed by my colleagues, but unfortunately I can't share results

Best Regards

Michal

onsdag den 2. september 2020 kl. 16.28.09 UTC+2 skrev snipped-for-privacy@highlandsniptechnology.com:

isolated and for hundreds of amps?

Looks like the hundreds of amps is transient current in a cap. Use the drop in a PCB trace or a bus bar. Transformer couple for isolation. All that would cost about a dollar.

I guessed pretty close on this one:

It measures 24.9 mOhms.

That would stand 100 amps for a reasonable transient. And I could have used much smaller resistors.

Seems like overkill to measure the current in a motor. In 50 ns.

They finally reveal the value of M in table 2. It has seven zeroes to the right of the decimal point. Curious notation.

One of the major features of the rogowski coil is its flexibility and its detachable physical termination, in portable instrumentation.

A printed circuit embodiment enforces physical limitations that, in turn, severely restricts its uses - but you can't beat the cost in a captive AC or low duty high current pulsed application, if you can get it to fit.

For captive applications, it's hard to beat the ACSxxx pcb-mounted, isolated sensors from Allegro, which will give you DC, as well, for continuous, asymmetrical and bidirectional currents.

RL

DC measurement is indeed a great feature, but there is a problem at the opposite end of the spectrum: the Hall effect sensors from Allegro are good to 1MHz max, which might not be sufficient in some applications. The Rogowski coil can easily go to 20MHz if one can afford a decent opamp. But admittedly, no DC response.

The open-loop Hall sensors from Allegro are also susceptible to external magnetic fields, down to DC (tested that the other day with ACS758). They are excellent if simplicity is at a premium. Still, if I had to actually *measure* something (either with accuracy or fast response in mind), I would first consider the shunt approach suggested by John. A fluxgate like the F03P050S05 would be my second choice, then a saturable reactor, then a Hall. Another nuisance with the Allegro sensors (and most others) is that the result is ratiometric, but they do their best not to tell you where the midpoint is, according to them. It kills differential interfacing instantly due to the unknown offset and temperature drift it introduces. Is that one additional pin cost-prohibitive?

There are magnetoresistive sensors as well, which can go to 2MHz:

but the resistance of their current path is about 10mOhms. That makes using them pointless, as a shunt of this value would allow me to measure the current with much better accuracy and speed. Not sure what problem they are trying to solve with that part.

Best regards, Piotr